Increase Index Limit about nri-elasticsearch HOT 3 CLOSED

@pzimno Thanks for the request -- I'm checking on whether this is somethign feasible. The cap was initially introduced because of limitations of the Insights API, so I can't promise we'll be able to increase it. However, there may be other ways to work around the limitation. I'll keep you updated 👍

from nri-elasticsearch.

What is the expected behavior when we reach out the index limit threshold? Should it drop the oldest info or panic?

I am seeing some failures related to that (error down below) using the given versions:

1. newrelic-infra-1.8.32
2. newrelic-infra-integrations-1.7.0
3. nri-elasticsearch-4.2.0

• Stacktrace:
  ERRO[0034] Integration command failed error="exit status 2" instance=elasticsearch integration=com.newrelic.elasticsearch prefix=config/elasticsearch stderr="[INFO] Collecting cluster metrics.\n[INFO] Collecting node metrics\n[INFO] Collecting common metrics.\n[INFO] Collecting indices metrics\n[ERR] Couldn't match index name in common index stats response: index 'staging-infra-aws_elasti_cache_memcached_cluster-v14' not contained in list (...) panic: runtime error: invalid memory address or nil pointer dereference\n[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x660d0a]\n\ngoroutine 1 [running]:\nmain.setIndicesStatsMetricsResponse(0xc4200e42c0, 0xc4203ea000, 0x518, 0x63f, 0xc4201de5a0, 0xc42017a0e0, 0x20, 0x0)\n\t/data/jenkins/infra-integrations-nri-elasticsearch-release-linux-package-3/go/src/github.com/newrelic/nri-elasticsearch/src/metrics.go:147 +0x14a\nmain.populateIndicesMetrics(0xc4200e42c0, 0x858de0, 0xc420098a40, 0xc4201de5a0, 0xc42017a0e0, 0x20, 0x0, 0x0)\n\t/data/jenkins/infra-integrations-nri-elasticsearch-release-linux-package-3/go/src/github.com/newrelic/nri-elasticsearch/src/metrics.go:107 +0x190\nmain.populateMetrics(0xc4200e42c0, 0x858de0, 0xc420098a40, 0x6bb341, 0x0)\n\t/data/jenkins/infra-integrations-nri-elasticsearch-release-linux-package-3/go/src/github.com/newrelic/nri-elasticsearch/src/metrics.go:35 +0x2b5\nmain.main()\n\t/data/jenkins/infra-integrations-nri-elasticsearch-release-linux-package-3/go/src/github.com/ne

• After the app panic we start receiving the following message:
  ERRO[0584] could not queue event component=Agent entityKey= error="Could not queue event: Queue is full." ERRO[0584] could not queue event component=Agent entityKey= error="Could not queue event: Queue is full."

from nri-elasticsearch.

This should be fixed in v4.3.2

from nri-elasticsearch.